Repro Capsule MCP enables you to create, analyze, and execute safe, reproducible bug reproduction cases with mandatory review, secret redaction, and export capabilities.
Create and manage cases: Start a new case with an automatically redacted bug report; list all persisted cases and retrieve full details (profile, artifacts, analyses, runs).
Inspect the project: Dynamically discover project manifests, available commands, environment variable names, and Git state within the case's project root.
Analyze evidence: Generate evidence-cited hypotheses to answer specific questions about the bug.
Add artifacts: Attach reviewed project files as redacted text or hash-only binary records (no unreviewed binary content).
Plan and execute commands: Prepare exact non-shell command plans with one-time approval tokens, declaring network and workspace‑write intents; execute with bounded output, timeouts, process‑group cleanup, and output redaction before persisting.
Compare runs: Diff two recorded execution transcripts by exit status, expectation match, and output hashes.
Export capsule: Package all case metadata and reviewed artifacts into a portable ZIP with a versioned manifest and reproduction guide; no secrets or tokens included.
Security: Automatic redaction of secrets, strict denial of unlisted commands, separate approvals for network and workspace write, allowed‑roots enforcement, and force‑cleanup of process groups on timeout.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Repro Capsule MCP start a case from this bug report"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
English | 한국어
Turn a bug report into a portable, reviewable reproduction capsule without silently copying a repository or executing model-generated shell text. Every report, command plan, transcript, artifact, and export stays evidence-linked and redacted.
flowchart LR
A[Bug report] --> B[Dynamic project profile]
B --> C[Evidence-based hypotheses]
B --> D[Reviewed command plan]
D --> E[One-time approval]
E --> F[Bounded execution]
C --> G[Portable capsule]
F --> GDynamic command discovery from current project manifests
Secret-redacted reports, text artifacts, stdout, and stderr
Binary artifacts represented by metadata and hashes instead of unreviewed content
Direct argv execution without a shell
One-time plan tokens with separate workspace-write and network approvals
Process-group timeout cleanup and bounded output capture
Run-to-run transcript comparison without guessing whether a fix is correct
ZIP export containing a versioned manifest and reproduction guide
Stdio and Streamable HTTP transports
Related MCP server: Soma MCP Server
uv tool install "git+https://github.com/efficjump/repro-capsule-mcp.git"
repro-capsule-mcp --transport stdioFor source development:
git clone https://github.com/efficjump/repro-capsule-mcp.git
cd repro-capsule-mcp
uv sync --all-extras --locked
uv run repro-capsule-mcp --transport stdio{
"mcpServers": {
"repro-capsule": {
"command": "repro-capsule-mcp",
"args": ["--transport", "stdio"]
}
}
}The installed command avoids embedding any local checkout path in client configuration.
Tool | Purpose |
| Create a bounded case with a redacted bug report |
| Browse persisted case state and evidence |
| Discover manifests, environment names, Git state, and commands |
| Produce host-model hypotheses from current evidence |
| Add reviewed text or hash-only binary evidence |
| Prepare an exact non-shell command and one-time token |
| Run one approved plan with bounds and redaction |
| Compare two execution transcripts |
| Create a portable ZIP and manifest |
Project roots must remain inside REPRO_CAPSULE_ALLOWED_ROOTS.
Sensitive filenames, symlinks, oversized files, and unreviewed binary content are excluded.
Commands not discovered from current project files are denied by default.
Network intent and workspace-write intent require separate approval.
Environment values and approval tokens are not written into capsule exports.
Every process uses a timeout, bounded output, and redaction before persistence.
This server is not an operating-system sandbox. Inspect and execute untrusted projects only in an isolated container or virtual machine.
repro-capsule-mcp --transport streamable-http --host 127.0.0.1 --port 8766The default endpoint is http://127.0.0.1:8766/mcp.
uv sync --all-extras --locked
uv run ruff format --check .
uv run ruff check .
uv run mypy src
uv run pytest --cov --cov-report=term-missing
uv buildSee architecture, security policy, and contribution guide.
Add one reviewed project file as redacted text or a binary hash-only record.
Inputs: case_id (required), relative_path (required), note, include_binary_content. Returns: object
Form evidence-cited hypotheses with host sampling, or return an explicit fallback.
Inputs: case_id (required), objective (required). Returns: object
Compare exit status, expectation match, and redacted output hashes for two runs.
Inputs: case_id (required), baseline_run_id (required), candidate_run_id (required). Returns: object
Consume one approval token and record bounded, redacted command output.
Inputs: case_id (required), plan_id (required), approval_token (required), allow_network, +1 more. Returns: object
Create a ZIP containing only persisted case metadata and reviewed artifact contents.
Inputs: case_id (required). Returns: object
List persisted reproduction cases without reading project files or running commands.
Inputs: no named parameters. Returns: result
Create a non-shell execution plan and one-time approval token from project evidence.
Inputs: case_id (required), purpose (required), argv, env_names, +5 more. Returns: object
Create a versioned case after constraining the project root and redacting the report.
Inputs: name (required), bug_report (required), project_root (required). Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 8 tools
Each tool maps to a distinct phase of the reproduction workflow: case creation, listing, analysis, artifact addition, command planning, execution, run comparison, and export. Even the closely related plan_command and execute_plan pair is clearly separated by planning versus consuming a token for execution.
All eight tool names follow a consistent verb_noun snake_case pattern such as start_case, add_artifact, execute_plan, and export_capsule. The action and target are predictable across the entire set.
Eight tools is a well-scoped size for a reproduction capsule server. Each tool provides a necessary capability in the lifecycle without redundancy or bloat.
The tool surface covers the core workflow end-to-end: create/list/analyze cases, add artifacts, plan and execute commands, compare runs, and export. Minor gaps exist around direct inspection or removal of individual artifacts/runs, but agents can work around these using the existing analysis and export tools.
Bug tracking for AI agents: session replay, console/network logs, repro packs, test plans, evidence.
Production bugs with replay, network, console and stack, ranked for your coding agent to fix.
Bounded authorization and independently verifiable execution evidence for agents.
Scoped agent execution. Server-side credentials, policy, budgets and verifiable receipts.